Commits on Feb 12, 2016
  1. Merge pull request #136 from aibor/updated_pragmas

    Updated the methods for pragma calls according to the currently
    committed Feb 12, 2016
Commits on Feb 10, 2016
  1. Makes private type-based pragma helper methods public.

    As discussed in #136, the type-based getter and setter helper methods
    for pragmas are public now. This allows direct usage in case there are
    no specific methods for a pragma.
    aibor committed Feb 10, 2016
  2. Adds missing helper methods for pragmas.

    Updated the methods for pragma calls according to the currently
    supported pragmas listed on The previous
    behaviour is maintained and extended by the new pragmas.
    Only thing changed is the auto_vacuum pragma, which has changed in the
    commit from 2007-04-26:
    So, the minimal version for this to work is SQLite 3.4.0
    aibor committed with aibor Jul 30, 2014
Commits on Feb 9, 2016
  1. Merge pull request #179 from tenderlove/rb_apply

    use rb_apply so we don't need to maintain two arrays
    committed Feb 9, 2016
  2. clean up test a little

    committed Feb 9, 2016
  3. fix warnings in tests

    committed Feb 9, 2016
  4. Switch to rb_ary_new2 and rb_ary_push to prevent GC segfault on params

    to user-defined function. The new test_function_gc_segfault is skipped
    because you don't want to see segfault when running rake test. Unskip the
    test and comment out rb_ary_push to see the test fail.
    jimflood committed Feb 9, 2016
  5. install gmp

    committed Feb 9, 2016
  6. add 2.3.0

    committed Feb 9, 2016
  7. Fix 1.8 support

    We need to check for libpthread when compiling.  Ruby >=1.9 already
    links against pthread, so it would Just Work (tm) but 1.8 would not.
    Considering how long this is broken with no one complaining, why do we still
    support 1.8?
    committed Feb 9, 2016
Commits on Feb 7, 2016
  1. Merge pull request #161 from franckverrot/bugfix/aggregate

    Align #create_aggregate with #create_aggregate_handler
    committed Feb 7, 2016
  2. Merge pull request #168 from masaedw/master

    Call close in ensure
    committed Feb 7, 2016
  3. Merge pull request #177 from eliskah/synopsis

    Added table creation that is necessary to run example in synopsis
    committed Feb 7, 2016
  4. Merge pull request #169 from jimflood/sqlite3_result_blob

    BLOB function result when ASCII-8BIT encoding
    committed Feb 7, 2016
Commits on Jan 22, 2016
  1. Added table creation that is necessary to run example in synopsis

    Fixed comment that describes creating a table
    eliskah committed Jan 22, 2016
Commits on Dec 5, 2015
Commits on Dec 4, 2015
  1. Fix warning from rake test

    jimflood committed Dec 4, 2015
Commits on Nov 5, 2015
Commits on Oct 29, 2015
  1. Call close in ensure

    timmuto committed Oct 29, 2015
Commits on Oct 10, 2015
  1. Prepare for release

    luislavena committed Oct 10, 2015
  2. Update ChangeLog with merges and details

    [skip ci]
    luislavena committed Oct 10, 2015
  3. Backport: release 1.3.10

    Backport release changes to be up-to-date with 1-3-stable.
    luislavena committed Oct 10, 2015
  4. CI: Test against all latest versions

    - Travis: Bump 2.1.x to latest
    - Travis: Test against 2.2.x
    - AppVeyor: Test against 2.1.x, 2.2.x
    luislavena committed Oct 10, 2015
  5. Merge pull request #159 from larskanis/add-rake-compiler-dock

    Make use of rake-compiler-dock for building windows binary gems
    luislavena committed Oct 10, 2015
Commits on Jul 16, 2015
  1. Add rake-compiler-dock for building Windows binary gems.

    rake-compiler-dock makes building Windows gems faster and with little to no setup
    compared to rake-compiler-dev-box. A simple 'rake gem:windows' should be enough.
    It is still possible to use a native cross build environment or rake-compiler-dev-box.
    larskanis committed with larskanis Jun 21, 2015
Commits on Jun 21, 2015
Commits on Jun 12, 2015
  1. Align #create_aggregate with #create_aggregate_handler

    Two issues are covered by this commit:
      * We make use of the value of `result` to define the value of the
        aggreator. Before that, `finalize` was used and this was not
        respecting the documentation;
      * Each call to `finalize` now resets the aggregator. In the future,
        there could be some kind of callback/DSL to make it configurable.
    Commit by:
      * Richard K. Michael <>
      * Franck Verrot <>
    franckverrot committed Jun 12, 2015